Properties of human connexin 31, which is implicated in hereditary dermatological disease and deafness

Abstract

The connexins are a family of at least 20 homologous proteins in humans that form aqueous channels connecting the interiors of coupled cells and mediating electrical and chemical communication. Mutations in the gene for human connexin 31 (hCx31) are associated with disorders of the skin and auditory system. Alterations in functional properties of Cx31 junctions are likely to play a role in these diseases; nonetheless, little is known about the properties of the wild-type channels. Here we show that hCx31 channels, like other connexin channels, are gated by voltage and close at low pH and when exposed to long-chain alkanols. Single-channel conductance of the fully open channel is approximately 85 pS, and it is permeable to Lucifer yellow, Alexa Fluor(350), ethidium bromide, and DAPI, which have valences of -2, -1, +1, and +2, respectively. In contrast to what has been reported for mouse Cx31, hCx31 appears to form functional heterotypic channels with all four connexins tested, Cx26, Cx30, Cx32, and Cx45. These findings provide an important first step in evaluating the pathogenesis of inherited human diseases associated with mutations in the gene for Cx31.

Fig. 1.

Expression and gating of hCx31. (A) Northern blot analysis of mRNA extracts from a Neuro2a cell line expressing hCx31 (left lane) and wild-type HeLa cells (right lane). (B) Western blot analysis of protein extracts from a Neuro2a cell line expressing hCx31 (left lane) and parental HeLa cells (right lane). (C) Immunostaining for hCx31WT expressed in HeLa cells shows punctate staining in a pattern consistent with the presence of gap junction plaques (arrows). (D) Representative current traces in response to V_j steps for homotypic hCx31 junctions between a pair of transfected Neuro2a cells. Both cells were voltage-clamped at 0 mV, junctional currents were recorded from cell 2, and cell 1 was stepped to voltages from zero to ±120 mV in increments of 20 mV. A 0.5-s +20-mV standardizing pulse step preceded each test pulse. (E) G_j–V_j relation for homotypic hCx31 channels between transfected Neuro2a cells. Each circle in the G_j–V_j plot is derived from a single determination of steady-state current. The solid line represents fits of the data to Boltzmann distributions for V_j of either sign. Parameters (+V_j, −V_j); G_min, 0.24, 0.21; G_max, 1.1, 1.1; A, 0.065, 0.076; V₀, 43.8, 43.9. (F) Chemical gating and pH sensitivity of hCx31 channels between HeLa cells. I_j was measured during application of repeated V_j ramps from −22 to +22 mV every 1.5 s. Application of 2 mM heptanol led to a rapid reduction in junctional current to below detectable levels. Junctional current rapidly recovered when heptanol was removed from the bath. Application of CO₂ saturated bath solution also caused rapid and reversible reduction in I_j.

Fig. 2.

Single-channel conductance and permeability of hCx31-EGFP channels measured in HeLa cells. (A) Junctional current records (middle trace) in response to voltage steps (bottom trace) from 0 to +70 mV. The dashed line on the current record is the residual conductance; the solid line is the fully open state of one channel. A point-by-point (1-ms) calculation of the junctional conductance is shown in the top row (g_j). The conductance of the fully open channel is ≈85 pS, and a predominant residual conductance of ≈15 pS is seen. Two Insets on the top show expanded records from regions defined by rectangles and illustrate the slow gating transition from the open state toward the closed state (Left Inset) and the fast gating transition from the substate to the open state (Right Inset). (B) Permeability of hCx31 junctions to fluorescent dyes. Cell 1 (asterisk) was loaded with the dye of interest via the patch pipette in the whole-cell recording configuration, and a gigaohm seal (on cell configuration) with cell 2 was established via a second patch pipette. In this way, the dye that crossed into cell 2 (the postjunctional cell) was not lost by diffusion into pipette 2. After a suitable interval (usually 5 min), the gigaohm seal on cell 2 was ruptured, g_j measured, and the presence of a cytoplasmic bridge excluded by application of 2 mM heptanol. As shown here, hCx31 is relatively nonselective, allowing passage of both negatively and positively charged dyes including Lucifer yellow (LY, −2), Alexa Fluor³⁵⁰ (AF³⁵⁰, −1), EtdBr (+1), and DAPI (+2).

Fig. 3.

hCx31 forms heterotypic junctions with Cx45. (A) The bottom trace shows the voltage applied to a HeLa cell expressing hCx31-EGFP and paired with a HeLa cell expressing Cx45 voltage clamped at 0. The middle trace shows the junctional currents measured in the cell expressing Cx45. The top trace shows g_j calculated point by point. These heterotypic junctions have a markedly asymmetric response to V_j. (B) Steady-state G_j–V_j relation for Cx45/Cx31-EGFP heterotypic junctions assembled from slow (4 min) ramps from 0 to +100 mV and from 0 to −100 mV. For increasing negativity on the hCx31-EGFP side, G_j increases up to ≈40 mV before decreasing again. For V_j of the opposite polarity, G_j decreases to a minimum at approximately −60 mV (Inset) and then shows a small increase. (C) Dependence of heterotypic Cx45/Cx31-EGFP coupling on holding current. A hCx31-EGFP-expressing cell (cell 1) was voltage-clamped at −10 mV and stepped by ±90 mV for 0.25 s with 0.25 s between pulses, and a coupled Cx45-expressing cell (cell 2) was current-clamped to approximately −12 mV. Hyperpolarizing voltage pulses starting before the record shown were applied to the hCx31-EGFP cell; these pulses caused g_j to increase, and there was a substantial response in cell 2. After ≈5 s, the pulses were changed to depolarizing, and the responses in cell 2 decreased gradually with successive pulses, indicating decrease in g_j. Expanded records in the Insets above show the changes associated with change in polarity in greater detail. Reapplication of hyperpolarizing pulses to cell 1 caused the responses in cell 2 to increase again, indicating increased g_j. At ≈30 s, the hyperpolarizing current in cell 2 was increased; the responses to hyperpolarizing steps in cell 1 remained about the same, but the responses to depolarizing steps were greatly reduced. At ≈58 s, the hyperpolarizing current in cell 2 was decreased below the initial value, and the responses to depolarization and hyperpolarization of cell 1 were of more nearly the same amplitude.

Fig. 4.

hCx31 forms heterotypic junctions with Cx32 and Cx30. Representative I_j traces in response to V_j steps (A and C) and average normalized G_j–V_j relations (B and D) for hCx31 paired heterotypically with Cx32 and Cx30. (A and C) Both cells were voltage-clamped at 0 mV, and cell 1 (expressing Cx32 or Cx30) was stepped to −20 mV for 0.2 s for normalization and then to voltages between +100 and −100 mV in 20-mV increments and I_j recorded from the hCx31-expressing cell. (B and D) Instantaneous (open triangles) and steady-state (filled squares) G_j–V_j relations for each cell pair. The dashed lines represent best fit straight lines for the instantaneous g_j). (E) An expanded view of the first 500 ms of traces in C. The dashed line is at I_j = 0.

Fig. 5.

hCx31 forms heterotypic junctions with Cx26. (A) I_j record (middle trace) from a hCx31-EGFP/Cx26 heterotypic cell pair in response to V_j ramps and steps applied to a cell expressing Cx26 (bottom trace). The corresponding junctional conductances are shown in the upper trace. I_j in response to a positive step declined to steady state over ≈15 s; fast g_j recovery measured during repeated V_j ramps after a positive V_j step indicates that the fast gating mechanism was responsible for this g_j decay. During a negative V_j step, g_j decayed instantaneously by ≈25% and remained relatively constant. The small ramps caused rapid changes in g_j (see B). (B) Changes in g_j induced by the small ramps decaying from +28 to −28 mV every 2 s; g_j increases quite linearly with voltage over that voltage range. The dashed line is a linear fit to the data. (C) Steady-state G_j–V_j relation for a hCx31-EGFP/Cx26 heterotypic cell pair assembled from slow (4 min) ramps from 0 to +100 mV and from 0 to −100 mV. The dashed line represents the linear fit of data points at V_js from −110 to +15 mV.